Interfaces

EvalInterface

Bases: object

The EvalInterface is an abstract base class that defines the interface for evaluation metrics in a generic evaluation system. It serves as a blueprint for concrete evaluation metric classes that implement specific evaluation logic.

Methods:

Name	Description
- calc	Abstract method that calculates the evaluation metric based on the provided scores, labels, and margins parameters. It returns an instance of a class that inherits from MetricInterface.
- get_name	Abstract method that returns the name of the evaluation metric. Concrete classes implementing this interface should provide their own implementation of this method.

calc abstractmethod

calc(scores, labels, margins)

Calculates the evaluation metric based on the provided scores, labels, and margins.

Parameters:

Name	Type	Description	Default
scores	ndarray	An array of anomaly scores.	required
labels	ndarray	An array of ground truth labels.	required
margins	[int(margin - before - anomaly), int(margin - after - anomaly)]	You can use margin to prune your evaluations if needed. Defined in https://github.com/dawnvince/EasyTSAD/blob/main/EasyTSAD/Controller/GlobalCfg.toml.	required

Returns:

Name	Type	Description
MetricInterface	Type[MetricInterface]	An instance of the event detection metric.

MetricInterface

Bases: object

The MetricInterface class is an abstract base class that defines the interface for metrics. It serves as a blueprint for creating subclasses that represent specific metrics.

The class includes three abstract methods: add(), avg(), and to_dict(). Subclasses inheriting from this class must implement these methods according to their specific metric calculations and requirements.

You should implement the following methods:

- add(self, other_metric): This abstract method represents the operation of combining two metrics. It takes another metric object (other_metric) as a parameter and is responsible for adding its values to the current metric.

- avg(self): This abstract method calculates the average value of the metric. It should be implemented by subclasses to compute the average based on the accumulated values.

- to_dict(self): This abstract method converts the metric object into a dictionary representation. It should return a dictionary containing the metric's values and any additional information needed for representation or storage.

add abstractmethod

add(other_metric)

This abstract method represents the operation of combining two metrics. It takes another metric object (other_metric) as a parameter and is responsible for adding its values to the current metric.

avg abstractmethod

avg()

This abstract method calculates the average value of the metric. It should be implemented by subclasses to compute the average based on the accumulated values.

to_dict abstractmethod

to_dict()

This abstract method converts the metric object into a dictionary representation. It should return a dictionary containing the metric's values and any additional information needed for representation or storage.